def test_fallback_ocr(self):
try:
# actually running tesseract takes ages -- for day-to-day
# testing we can just as well use the canned hocr.html
# files that _copy_sample fixes for us.
if not os.environ.get("FERENDA_TEST_TESSERACT"):
raise errors.ExternalCommandError
reader = PDFReader(
filename="test/files/pdfreader/scanned-ecma-99.pdf",
workdir=self.datadir,
images=False)
except errors.ExternalCommandError:
self._copy_sample()
reader = PDFReader(
filename="test/files/pdfreader/scanned-ecma-99.pdf",
workdir=self.datadir,
images=False)
self.assertTrue(reader.is_empty())
reader = PDFReader(filename="test/files/pdfreader/scanned-ecma-99.pdf",
workdir=self.datadir,
ocr_lang="eng")
self.assertFalse(reader.is_empty())
self.assertEqual(2, len(reader))
self.assertEqual("EUROPEAN COMPUTER MANUFACTURERS ASSOCIATION",
util.normalize_space(str(reader[0][1])))
def test_fallback_ocr(self):
try:
# actually running tesseract takes ages -- for day-to-day
# testing we can just as well use the canned hocr.html
# files that _copy_sample fixes for us.
if not os.environ.get("FERENDA_TEST_TESSERACT"):
raise errors.ExternalCommandError
reader = PDFReader(filename="test/files/pdfreader/scanned-ecma-99.pdf",
workdir=self.datadir,
images=False)
except errors.ExternalCommandError:
self._copy_sample()
reader = PDFReader(filename="test/files/pdfreader/scanned-ecma-99.pdf",
workdir=self.datadir,
images=False)
self.assertTrue(reader.is_empty())
reader = PDFReader(filename="test/files/pdfreader/scanned-ecma-99.pdf",
workdir=self.datadir,
ocr_lang="eng")
self.assertFalse(reader.is_empty())
self.assertEqual(2, len(reader))
self.assertEqual("EUROPEAN COMPUTER MANUFACTURERS ASSOCIATION",
util.normalize_space(str(reader[0][1])))
def test_basic(self):
try:
reader = PDFReader(filename="test/files/pdfreader/sample.pdf",
workdir=self.datadir)
except errors.ExternalCommandError:
self._copy_sample()
reader = PDFReader(filename="test/files/pdfreader/sample.pdf",
workdir=self.datadir)
# a temporary copy of the pdf file should not be lying around
# in workdir
# print("Checking if %s has been unlinked" % (self.datadir +
# os.sep + "sample.pdf"))
self.assertFalse(os.path.exists(self.datadir + os.sep + "sample.pdf"))
# but the XML file should be stored for subsequent parses
self.assertTrue(os.path.exists(self.datadir + os.sep + "sample.xml"))
# The PDF contained actual textboxes
self.assertFalse(reader.is_empty())
self.assertEqual(len(reader), 1)
# first page, first box
title = str(reader[0][0])
self.assertEqual("Document title ", title)
self.assertEqual(570, reader.median_box_width())
page = reader[0]
self.assertEqual(
"Page 1 (892 x 1263): 'Document title This is a simple documen...'",
str(page))
# an uncropped doc should have nine nonempty textboxes
self.assertEqual(9, len(list(page.boundingbox())))
# a smaller bounding box yields just one
self.assertEqual(1, len(list(page.boundingbox(190, 130, 230, 460))))
# cropping it with the same dimensions
# NOTE: This will fail if convert (from imagemagick) isn't installed)
try:
page.crop(190, 130, 230, 460)
except errors.ExternalCommandError:
# the rest of the tests cannot succeed now. FIXME: We
# should try to find a way to run them anyway
return
# should also result in just one box -- the bottom one
boxes = list(page.boundingbox())
self.assertEqual(1, len(boxes))
box = boxes[0]
self.assertEqual("This is a simple document in PDF format. ", str(box))
self.assertEqual('#000000', box.font.color)
self.assertEqual(16, box.font.size)
self.assertEqual('1', box.font.id)
self.assertEqual('Cambria', box.font.family)
# this box should have four text elements
self.assertEqual(4, len(box))
self.assertEqual(None, box[0].tag)
self.assertEqual("i", box[1].tag)
self.assertEqual("ib", box[2].tag)
self.assertEqual(None, box[3].tag)
def test_basic(self):
try:
reader = PDFReader(filename="test/files/pdfreader/sample.pdf",
workdir=self.datadir)
except errors.ExternalCommandError:
self._copy_sample()
reader = PDFReader(filename="test/files/pdfreader/sample.pdf",
workdir=self.datadir)
# a temporary copy of the pdf file should not be lying around
# in workdir
# print("Checking if %s has been unlinked" % (self.datadir +
# os.sep + "sample.pdf"))
self.assertFalse(os.path.exists(self.datadir + os.sep + "sample.pdf"))
# but the XML file should be stored for subsequent parses
self.assertTrue(os.path.exists(self.datadir + os.sep + "sample.xml"))
# The PDF contained actual textboxes
self.assertFalse(reader.is_empty())
self.assertEqual(len(reader), 1)
# first page, first box
title = str(reader[0][0])
self.assertEqual("Document title ", title)
self.assertEqual(570, reader.median_box_width())
page = reader[0]
self.assertEqual("Page 1 (892 x 1263): 'Document title This is a simple documen...'", str(page))
# an uncropped doc should have nine nonempty textboxes
self.assertEqual(9, len(list(page.boundingbox())))
# a smaller bounding box yields just one
self.assertEqual(1,
len(list(page.boundingbox(190, 130, 230, 460))))
# cropping it with the same dimensions
# NOTE: This will fail if convert (from imagemagick) isn't installed)
try:
page.crop(190, 130, 230, 460)
except errors.ExternalCommandError:
# the rest of the tests cannot succeed now. FIXME: We
# should try to find a way to run them anyway
return
# should also result in just one box -- the bottom one
boxes = list(page.boundingbox())
self.assertEqual(1, len(boxes))
box = boxes[0]
self.assertEqual("This is a simple document in PDF format. ", str(box))
self.assertEqual('#000000', box.font.color)
self.assertEqual(16, box.font.size)
self.assertEqual('1', box.font.id)
self.assertEqual('Cambria', box.font.family)
# this box should have four text elements
self.assertEqual(4, len(box))
self.assertEqual(None, box[0].tag)
self.assertEqual("i", box[1].tag)
self.assertEqual("ib", box[2].tag)
self.assertEqual(None, box[3].tag)
